Smollm2 Prompt Enhance GGUF

🚀 Transformers Library

This library is designed for text generation, leveraging techniques like TRL and SFT.

🚀 Quick Start

The transformers library is a powerful tool for natural language processing. It allows you to easily load pre - trained models and perform various tasks such as text generation.

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

device = "cuda" if torch.cuda.is_available() else "cpu"
model_id = "gokaygokay/SmolLM2-Prompt-Enhance"
tokenizer_id = "HuggingFaceTB/SmolLM2-135M-Instruct"
# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained(tokenizer_id )
model = AutoModelForCausalLM.from_pretrained(model_id).to(device)

# Model response generation functions
def generate_response(model, tokenizer, instruction, device="cpu"):
    """Generate a response from the model based on an instruction."""
    messages = [{"role": "user", "content": instruction}]
    input_text = tokenizer.apply_chat_template(
        messages, tokenize=False, add_generation_prompt=True
    )
    inputs = tokenizer.encode(input_text, return_tensors="pt").to(device)
    outputs = model.generate(
        inputs, max_new_tokens=256, repetition_penalty=1.2
    )
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return response

def print_response(response):
    """Print the model's response."""
    print(f"Model response:")
    print(response.split("assistant\n")[-1])
    print("-" * 100)

prompt = "cat"

response = generate_response(model, tokenizer, prompt, device)
print_response(response)

# a gray cat with white fur and black eyes is in the center of an open window on a concrete floor. 
# The front wall has two large windows that have light grey frames behind them. 
# here is a small wooden door to the left side of the frame at the bottom right corner. 
# A metal fence runs along both sides of the image from top down towards the middle ground.
# Behind the cats face away toward the camera's view it appears as if there is another cat sitting next to the one 
# they're facing forward against the glass surface above their head.

💻 Usage Examples

Basic Usage

# The basic usage of generating a response based on a prompt
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

device = "cuda" if torch.cuda.is_available() else "cpu"
model_id = "gokaygokay/SmolLM2-Prompt-Enhance"
tokenizer_id = "HuggingFaceTB/SmolLM2-135M-Instruct"
tokenizer = AutoTokenizer.from_pretrained(tokenizer_id )
model = AutoModelForCausalLM.from_pretrained(model_id).to(device)

prompt = "cat"
response = generate_response(model, tokenizer, prompt, device)
print_response(response)

📚 Documentation

Training Script

You can access the training script through this link.

📄 License

This project is licensed under the Apache - 2.0 license.

📦 Information Table